FreeNAS Build with 10GBe and Ryzen

rvassar

Guru
Joined
May 2, 2018
Messages
971
Well, that part's good. Now if you can find a friend who's a motherboard manufacturer engineer and ask about whether or not logging and reporting is properly implemented, then we might have a complete story :) ;-)

Agreed. But I suspect the odds are pretty low. Lots of chip outfits here, but all the systems stuff orbits Dell, and they're not into AMD. It's almost a Longhorn vs. A&M style grudge. ;)

But I found it interesting that he'd text me an answer like that 4+ days after I asked. I'm hoping he can get me some documentation indicating a change in support status, and maybe a list of supporting chipsets. That would be actually useful, and could be thrown at board vendors... I'll keep working on it.
 
Last edited:

rvassar

Guru
Joined
May 2, 2018
Messages
971
Ok... So the support is there in the chips, but they foist the validation off on the motherboard manufacturers. The motherboard manufacturers only validate ECC with the Ryzen Pro chips, which you never see for sale. They're only available in bulk tray, they're not retail packaged.

So you should be able to get away with a retail packaged chip, but it won't be validated, and your motherboard manufacturer will balk at supporting you.
 

Mastakilla

Patron
Joined
Jul 18, 2019
Messages
202
Ok... So the support is there in the chips, but they foist the validation off on the motherboard manufacturers. The motherboard manufacturers only validate ECC with the Ryzen Pro chips, which you never see for sale. They're only available in bulk tray, they're not retail packaged.

So you should be able to get away with a retail packaged chip, but it won't be validated, and your motherboard manufacturer will balk at supporting you.
That sounds like it should be enough for a home-NAS for me... It's not that I'm running business critical stuff on it ;)

Hope someone can take a look at my (non?-)SLOG-understanding now (previous page) :D Thanks!
 

rvassar

Guru
Joined
May 2, 2018
Messages
971
That sounds like it should be enough for a home-NAS for me... It's not that I'm running business critical stuff on it ;)

Or pick up Pro chip via eBay... I'm looking at one listing right now.

Hope someone can take a look at my (non?-)SLOG-understanding now (previous page) :D Thanks!

L2ARC.... You're always better off just adding more RAM. If you can't, and have other workloads (jails/vm's), move the workloads to free up the RAM.

SLOG - NFS writes are per the RFC's mandatory O_SYNC. The problem with the smaller Optane's is they're not all that fast, and in a 10GbE application they may actually become the bottleneck. The neat thing about FreeNAS is you can turn the sync behavior off at the dataset layer. So I have a pool with NFS shares that are RFC compliant, and a couple for VM datastore's that are not.
 

Mastakilla

Patron
Joined
Jul 18, 2019
Messages
202
Ok in meantime I've bought the hardware in my signature and did a default FreeNAS install. Will read some more and try to fine-tune it some more.
I also still have to buy the 10Gbit NIC for my desktop, so currently it is still attached with useless 1Gbit :eek:
Waiting for a good deal on a 2nd hand x550 NIC...

Well, that part's good. Now if you can find a friend who's a motherboard manufacturer engineer and ask about whether or not logging and reporting is properly implemented, then we might have a complete story :) ;-)
https://www.reddit.com/r/Amd/comments/93g9j9/guide_how_to_confirm_unbuffered_unregistered_ecc/
I'll try the test-instructions from the above link later to confirm...

SLOG - NFS writes are per the RFC's mandatory O_SYNC. The problem with the smaller Optane's is they're not all that fast, and in a 10GbE application they may actually become the bottleneck. The neat thing about FreeNAS is you can turn the sync behavior off at the dataset layer. So I have a pool with NFS shares that are RFC compliant, and a couple for VM datastore's that are not.
I still don't really understand this I'm afraid. Are you saying that sync isn't "required" (and is not beneficial?) for NFS shares? And that it is mainly "required" (and beneficial?) for datasets that you're using as VM datastore?
I currently only have a dataset as SMB share for Windows and no VMs (yet). I'll probably deploy things like Plex and qBitTorrent later. But I'll try to have most of VMs local on SSD instead of NAS (don't really see the benefit in that yet). So does that mean that a SLOG is useless for me?

IF I would need a SLOG, then I'm currently considering the Intel Optane M15 64GB. It has power failure protection build-in, good enough performance (maxes out my mobo M.2 PCIe 3.0 x2 / PCIe 2.0 x4 with 2GB/sec). The only problem is that I haven't found it for sale yet... ;) Should be soon though...
 

Mastakilla

Patron
Joined
Jul 18, 2019
Messages
202
For those having trouble with a similar built, here is how I "kinda" did it...
Please keep in mind that I didn't log everything I did, so this is from memory and perhaps I'm forgetting something...

So, this is my first built with IPMI. I didn't install a graphics card, but I was hoping it would work with just a network to the IPMI (and it did :) ).
  • I installed all hardware except the HBA / HDDs in the case
  • I attached a monitor to the D-Sub connector of the mobo (I think this is of the IPMI?)
  • I attached a USB keyboard / mouse
  • I attached a networkcable between the IPMI NIC and the router
  • In my router-admin-page I found the IP that DHCP of the router assigned to the IPMI
  • I browsed to the IPMI web page and logged in with the default credentials (admin / admin)
  • I configured both router and IPMI to use a static IP and changed the admin password
  • I updated the IPMI to v1.40.00 using 'maintenance' - 'firmware update' in the IPMI
  • I update the BIOS to v3.20 using 'maintenance' - 'BIOS Update' in the IPMI
    • I think I used the 'Instant Flash' download for this
    • Yes, this worked, even though my CPU wasn't supported by the older BIOS that was installed
    • No, this isn't properly documented anywhere that I could find
  • I tried opening 'Remote Control' and booting, but didn't get any visual and many beeps / error codes
  • In the end, I think unplugging both the monitor and USB keyboard / mouse, allowed me to finally boot and get an visual in 'Remote Control'
  • At one point I also tried disabling 'Onboard Graphics' in the BIOS (to save some power usage), but that also caused it to no longer boot. Will try fine tuning the BIOS a bit more later...
  • Then I installed the HBA / HDDs
  • The HDDs support the enterprise feature PWDIS (Power Disable). The PSU does not support this. So none of the HDDs would power up.
    • I fixed this by disconnecting the 3.3v from the modular SATA power cable on the PSU side, using the below method:
      https://www.youtube.com/watch?v=kXMwCQ9fIVM
    • For 1 of 2 cable this succeeded without any damage to the disconnected cable, for the other cable, 1 of the 2 hooks got bended / forced (I do think I'll manage to re-attach both without anyone being able to tell, in case I would ever need to use warranty)
  • The HBA was already in IT mode, but was running an old firmware. Upgrading it was a struggle...
    • Bootable (Free)DOS USB stick didn't work as my BIOS is UEFI ("Failed to initialize PAL" error)
    • Update using EFI shell also refused to work, even though I was using the same old 'Shell_Full.efi' file that others used succesfully (I kept on getting “InitShellApp: Application not started from Shell” error)
    • In the end I just installed Windows 10 and used the Windows x64 installer! :D
  • Finally I installed FreeNAS 11.2 U6. No issues occured during the install.
  • Both Windows 10 and FreeNAS, I installed using 'CD Image' functionality inside the 'Remote Control' of the IPMI. It's probably slower, but saves you from having to create and mess with USB sticks.
 
Last edited:

Mastakilla

Patron
Joined
Jul 18, 2019
Messages
202
...
IF I would need a SLOG, then I'm currently considering the Intel Optane M15 64GB. It has power failure protection build-in, good enough performance (maxes out my mobo M.2 PCIe 3.0 x2 / PCIe 2.0 x4 with 2GB/sec). The only problem is that I haven't found it for sale yet... ;) Should be soon though...
https://www.anandtech.com/show/14903/intel-shares-new-optane-and-3d-nand-roadmap
Notably missing from Intel's presentation is anything new for consumers based on Optane. They are still touting the Optane Memory H10 two-in-one SSD, but the second-generation consumer Optane M.2 drives (Optane Memory M15 and Optane SSD 815P) have been officially cancelled. The 815P is dead because 118GB is simply too small for a standalone SSD to be competitive, and the M15 is dead because there aren't enough systems still shipping with mechanical hard drives to support another generation of cache drives. The current Optane Memory M10 and Optane SSD 800P are not being discontinued yet.
There goes my plan... :(
 

rvassar

Guru
Joined
May 2, 2018
Messages
971
I still don't really understand this I'm afraid. Are you saying that sync isn't "required" (and is not beneficial?) for NFS shares? And that it is mainly "required" (and beneficial?) for datasets that you're using as VM datastore?
I currently only have a dataset as SMB share for Windows and no VMs (yet). I'll probably deploy things like Plex and qBitTorrent later. But I'll try to have most of VMs local on SSD instead of NAS (don't really see the benefit in that yet). So does that mean that a SLOG is useless for me?

IF I would need a SLOG, then I'm currently considering the Intel Optane M15 64GB. It has power failure protection build-in, good enough performance (maxes out my mobo M.2 PCIe 3.0 x2 / PCIe 2.0 x4 with 2GB/sec). The only problem is that I haven't found it for sale yet... ;) Should be soon though...

The NFS protocol itself mandates sync in the engineering documents that define the protocol. (See: RFC 3530 pg. 15 ) The file handle option flag for it in C/C++ is "O_SYNC", sorry I was unclear. FreeNAS allows you to turn this behavior off as an option. The problem with a SLOG that doesn't have power failure protection is it may confirm a write has occurred and then loose the data in a power event. In that case, having sync behavior enabled on FreeNAS doesn't really offer you any protection, so you might as well turn it off rather than run a non-PFP SLOG device. As for my situation... I have a collection of VM's that I'm not particularly worried about from a file corruption standpoint. I have backups, and snapshots, and ESXi's VMFS doesn't seem to corrupt all that easily, so rather than find a PFP compliant SLOG, I turned sync off for that datastore. Other datastore's & shares in the same ZFS pool I do want the full compliance and protection, and it is enabled. I keep some VM's on the NAS, as "poor man's vmotion", and some on a local SSD.
 

Mastakilla

Patron
Joined
Jul 18, 2019
Messages
202
Thanks for the clarification!

I guess then if I don't want to have a downgrade when replacing my 10 year old hardware RAID5 (incl battery) with FreeNAS, I'll need to leave sync enabled and only install SLOG if it has powerfailure protection :/

Too bad the M15 got cancelled, cause the 900p is still very pricy... :(
 

Wansteen

Cadet
Joined
Oct 14, 2019
Messages
4
hey mastakilla
are you still happy with your setup?
wanted to build a similar setup, as soon as the mainboard gets down to a reasonable price again...(rised from 300 to 400 € here -.-)

did you test your power consumption?
 

Mastakilla

Patron
Joined
Jul 18, 2019
Messages
202
Hi Wansteen. That is a bit early for me to say... I did get FreeNAS installed without big showstoppers, but I didn't spend enough time with FreeNAS to know well how it behaves.
Currently I'm actually messing around with Windows 10 and CentOS for overclocking the memory and testing the ECC capability.
Getting Linux to boot has been quite a struggle. Most flavors besides CentOS and RHEL don't work at all. Today I've received a beta BIOS from Asrock which might solve those issues though (it finally brings the BIOS to 1.0.0.3 ABBA). I'm going to test it now...

For more info on the Mobo itself and the issues people are having with it, look here:
https://forum.level1techs.com/t/asrock-rack-has-created-the-first-am4-socket-server-board/139490/535
https://forum.level1techs.com/t/asrock-rack-x470d4u2-2t/147588/23

As always, keep in mind that people usually only speak when they're having issues... So good experiences may be harder to find written down ;)

Anyway, so far, I'm not yet unhappy ;) Depending on how those new BIOS work, that may change (for better or worse)...
 

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,175
it finally brings the BIOS to 1.0.0.3 ABBA). I'm going to test it now...
So I guess the roar of cannons is coming from Intel's camp...
 

Mastakilla

Patron
Joined
Jul 18, 2019
Messages
202
After a long silence, it is about time for me to report back here. I haven't been sitting still at all. I've been trying very hard to validate ECC functionality of my hardware.

I've used many different sources on the internet for figuring this out. Below list are only a couple, that I still have open in my browser:
https://hardwarecanucks.com/cpu-motherboard/ecc-memory-amds-ryzen-deep-dive/
https://hardwarecanucks.com/forum/t...ryzen-a-deep-dive-comment-thread.75041/page-4
https://forum.level1techs.com/t/asrock-rack-x470d4u2-2t/147588/52
https://forum.level1techs.com/t/asr...-server-boards-x470d4u-x470d4u2-2t/139490/846

Detection of ECC support
Here I tried to figure out how various OS / software detect the ECC capabilities of my hardware.

IPMI
The IPMI Event Log contains sensors for "DRAM ECC Error" events, which can trigger actions like sending me an email.
1576835509345.png


BIOS
The BIOS is has a lot of options, both consumer grade options (like all kinds of OC options) and server grade options (like IPMI / ECC related options). I don't know
1576835613256.png
1576835645530.png
1576835812600.png

Except for trying to disable it for testing (as shown below), I didn't really play much with these yet and I also don't really have a good understanding of most of these options. Please feel free to teach me, if it is relevant for my "search".

MemTest86
MemTest86 latest version doesn't know about the ECC capability of my hardware yet.
1576836064540.png

When asked about it (1 month ago), this was Passmark their response:
Hello,

It appears that ECC detection support for Ryzen 3000 series chipset has not yet been implemented in MemTest86.

We are in contact with our AMD representative at the moment to obtain access to the necessary datasheets. Once acquired, we should be able to implement support for your chipset.

I should add that AMD haven't been very responsive recently.
We asked this question to AMD about 2 months back, but nothing yet.

Will keep you updated.

Kind regards,
Richard Ng

Linux
Linux requires kernel 5.4 to fully support Ryzen 3000 CPUs, so I've used the testing-release Fedora Rawhide for below findings. It does not work with kernel 5.3 or older (so the latest stable Ubuntu does not support it for example)!
1576838395763.png

It says "DRAM ECC enabled" and "using x16 syndromes", which is good.

Windows
I've used the latest Windows (first 1903, later I've also tried with 1909) with the latest AMD chipset drivers (first amd_software_1.09.27.1033.zip, later amd_chipset_software_1.11.22.454.zip).
1576838921572.png
(Windows itself)
6 = multi-bit ECC
TotalWidth is greater than DataWidth, so that is good, although some say it should be 72 instead of 128.
1576838954346.png
(CPU-Z)
1576839040175.png

1576839052719.png


FreeNAS / FreeBSD
I'm not sure how to properly see it in FreeNAS / FreeBSD. Someone on the FreeBSD forum tried to figure out the same with FreeBSD 12.1, but also didn't really find an answer:
https://forums.freebsd.org/threads/how-to-find-out-if-ecc-is-enabled.72839/

It wouldn't surprise me if it is similar to Linux and that full support for Ryzen 3000 needs to be added through the FreeBSD kernel first. As on Linux support was only added since a couple of weeks (kernel 5.4), it wouldn't surprise me if support hasn't been added yet to FreeBSD at all.
So if it doesn't even exist yet on FreeBSD, I guess I'll need to wait like forever before FreeNAS will have it... :(

Can someone perhaps confirm if this assumption is correct?

Here is my result in FreeNAS 11.2:
root@FreeNAS[~]# dmidecode -t memory
# dmidecode 3.1
# SMBIOS entry point at 0xed318000
Found SMBIOS entry point in EFI, reading table from /dev/mem.
SMBIOS 3.2 present.
# SMBIOS implementations newer than version 3.1.1 are not
# fully supported by this version of dmidecode.

Handle 0x000F, DMI type 16, 23 bytes
Physical Memory Array
Location: System Board Or Motherboard
Use: System Memory
Error Correction Type: Multi-bit ECC
Maximum Capacity: 128 GB
Error Information Handle: 0x000E
Number Of Devices: 4

Handle 0x0019, DMI type 17, 84 bytes
Memory Device
Array Handle: 0x000F
Error Information Handle: 0x0018
Total Width: 128 bits
Data Width: 64 bits
Size: 16384 MB
Form Factor: DIMM
Set: None
Locator: DIMM 1
Bank Locator: P0 CHANNEL A
Type: DDR4
Type Detail: Synchronous Unbuffered (Unregistered)
Speed: 2666 MT/s
Manufacturer: Kingston
Serial Number: F806A751
Asset Tag: Not Specified
Part Number: 9965745-002.A00G
Rank: 2
Configured Clock Speed: Unknown
Minimum Voltage: 1.2 V
Maximum Voltage: 1.2 V
Configured Voltage: 1.2 V

Handle 0x001E, DMI type 17, 84 bytes
Memory Device
Array Handle: 0x000F
Error Information Handle: 0x001D
Total Width: 128 bits
Data Width: 64 bits
Size: 16384 MB
Form Factor: DIMM
Set: None
Locator: DIMM 1
Bank Locator: P0 CHANNEL B
Type: DDR4
Type Detail: Synchronous Unbuffered (Unregistered)
Speed: 2666 MT/s
Manufacturer: Kingston
Serial Number: FA069F51
Asset Tag: Not Specified
Part Number: 9965745-002.A00G
Rank: 2
Configured Clock Speed: Unknown
Minimum Voltage: 1.2 V
Maximum Voltage: 1.2 V
Configured Voltage: 1.2 V

root@FreeNAS[~]#
 
Last edited:

Mastakilla

Patron
Joined
Jul 18, 2019
Messages
202
Actual testing of ECC reporting / logging
As Jgreco mentioned earlier in this thread, the actual error correction is only a part of ECC functionality. I now understand that reporting / logging the occurance of corrected single-bit-errors and uncorrected multi-bit-errors is at least as important, because you need to know if a DIMM is dying, so you can replace it before it causes data corruption.
So that is why I didn't stop here and I've also tried to trigger memory errors and see if they are correctly handled.

In the beginning I had a lot of issues with achieving unstable, but bootable, memory settings. Either it would be stable or it wouldn't boot at all (and require a CMOS reset). After a lot testing, the trick was to find the highest bootable setting and then lowering the voltage until it became unstable.

And a lot more testing I've done since then :)

To give you an idea, here's the Excel I've used to keep track of my testing:
1576843257518.png


I’ve tested from hardly bootable to slightly unstable using MemTest86, memtester (on Fedora Rawhide with kernel 5.4.0.0.rc3 and 5.4.0.2) and prime95/aida64_bench/Ryzen_Master_test (on a fully updated Windows 10 Pro, first with amd_software_1.09.27.1033.zip and later with amd_chipset_software_1.11.22.454.zip chipset drivers).

In mean time, I’ve had millions of memory errors (in total) in very varied conditions. It seems almost impossible to me if there was not a single single-bit-error or two-bit-error in all these millions of errors.

But… Unfortunately I couldn’t find any report of a corrected or logged memory error in either the IPMI Event Log, the Linux edac-util or the Windows Event Viewer (even though all of these report ECC to be active and correctly configured - see my posts above).
1576843947657.png

1576843984520.png

1576844029109.png

1576844042383.png


Now I know that doesn’t mean that no single-bit-error memory error-corrections have happened, but that is only half of what ECC functionality is. Reporting / logging these memory error-corrections is at least as important as the actual correcting itself (How else can you know your RAM is dying or is unstable).

So it seems to me that ECC is not working correctly on this motherboard with a Ryzen 3000 CPU (I don't have the older Ryzen CPUs for testing).

I've reported this to Asrock Rack and they've send me the following response:

Dear Mastakilla,

Due to X470 belongs to desktop series
It’s not like server MB has native support of ECC report.
We are checking with RD and AMD if X470 can support ECC report.
We will reply to you ASAP

Best regards,
Kevin
Asrock Rack Incorporation

I've replied to this with:

Hi Kevin,

Thanks a lot for looking into this! That is greatly appreciated…

I understand that the X470 is indeed a desktop chipset. Also all AM4 CPUs don’t have officially validated ECC support by AMD (although AMD confirmed that it wasn’t disabled).
So you could argue that non-validated half-working (not reporting / logging) ECC support is acceptable. And I also agree with that, for consumer brands like Asrock, Asus, MSI, etc.
However, if a brand like Asrock Rack or SuperMicro creates a X470 motherboard with “Supports 4x DDR4 ECC and non-ECC UDIMM, max. 128 GB” in the specifications and if the IPMI Event Log contains sensors for “DRAM ECC Error A1/A2/B1/B2”, then people (like myself) will assume that it is actually working and validated. In that case, I don’t think that it is acceptable for it not to work 100%, as people buying these brands, actually are expecting it to fully work. I don’t think that is a reputation or name you are looking for, as a brand called “Asrock Rack”

Please let me know if there is anything else I can do to assist.

Kind regards,

Mastakilla

The response from Asrock Rack seems to admit that it currently does not fully support ECC, however, it could also just mean that Kevin is not sure about it... So I'm hoping for a decent response from their R&D.

FreeNAS specific questions
Now I do wonder what you FreeNAS gurus consider as the correct / desired way of how ECC should work...

Wendell (owner of Level1Techs.com) explained me there are couple ways ECC can be implemented (in quotes below):

Platform First
"The hardware does what it can to recover from the error silently. It’s not necessary logged."
I suppose this would be if the errors are logged in the IPMI Event Log, which could then send an email to me to report the issue. In that case, I suppose, it isn't even required that FreeNAS or FreeBSD has support for Ryzen 3000 in its kernel, as it will work just as good without it. I do find it strange that he said "It’s not necessary logged". I've send him a follow up question on that...

OS First
"The error is forwarded to the os for it to decide what to do and the hardware does not halt, panic or send a non maskable interrupt."
For this properly work, I guess you need a couple things:
  • Full Ryzen 3000 support in the kernel (I guess this could take some time)
  • Proper handling of the discovered memory errors by FreeNAS, so that the user is notified of them. I didn't really find yet if this properly set by default or if specific user configuration is required to make this work.

I saw that Asus and Gigabyte both have x570 mobos with explicit ECC support. Gigabyte explicitly mentions Ryzen 3000 even, Asus is less clear on that. But as these Mobos don't have an IPMI, I suppose that there must be "OS First" instead of "Platform First", and as far as I could tell, that is no solution for FreeNAS as the OS probably currently does not fully support the Ryzen 3000 yet...
 
Last edited:

Ericloewe

Server Wrangler
Moderator
Joined
Feb 15, 2014
Messages
20,175
Platform First
"The hardware does what it can to recover from the error silently. It’s not necessary logged."
I suppose this would be if the errors are logged in the IPMI Event Log, which could then send an email to me to report the issue. In that case, I suppose, it isn't even required that FreeNAS or FreeBSD has support for Ryzen 3000 in its kernel, as it will work just as good without it. I do find it strange that he said "It’s not necessary logged". I've send him a follow up question on that...
Bryan Cantrill mentions in one of his BSDnow interviews (they total several hours, so I can't tell you which one) that an unspecified vendor told him they just hide ECC errors and that, when asked, the reasoning was something like the error rate was too insane to warn on.
 
Joined
May 10, 2017
Messages
838
Earlier this year I got two ECC corrected warning errors in the same day, separated buy a few hours, on a Supermicro X9SCL-F, never got one before and none since, and I have multiple similar boards.

localhost *Warning* Memory - Correctable memory error ; OEM Event Data3 code = 00h

Maybe one of those cosmic ray things ?
 

jgreco

Resident Grinch
Joined
May 29, 2011
Messages
18,681
Earlier this year I got two ECC corrected warning errors in the same day, separated buy a few hours, on a Supermicro X9SCL-F, never got one before and none since, and I have multiple similar boards.

localhost *Warning* Memory - Correctable memory error ; OEM Event Data3 code = 00h

Maybe one of those cosmic ray things ?

Yup. Seen that kind of thing in our systems too. I figure Doctor Manhattan must have been passing by.
 

Mastakilla

Patron
Joined
Jul 18, 2019
Messages
202
I have some good news, which should make reproducing (and validating after fixing) the issue a lot easier for Asrock Rack! It seems like Passmark have updated their MemTest86 product from version 8.2, which didn’t fully support Ryzen 3000, to version 8.3, which does fully support Ryzen 3000 (they forgot to put it in the changelog though).

This is very interesting, as MemTest86 Pro (not the Free version) supports ECC Injection:

“ECC injection: Enabled/Disabled (Pro version only) - if CC detection/correction is supported/enabled and CC injecton is supported by the system this option enables/disables injecton of CC errors to simulate how the system responds to real CC errors. CC errors are injected at the start of each individual test. If CC injection is successful the details of the CC error shall be reported and displayed on screen as if an actual CC error was detected.

Notes Although ECC injection may be supported by your hardware, it may be locked by the BIOS. Some BIOS may allot you to unlock the ECC injection feature in the BIOS setup.”

And Asrock Rack did do very well on that regard, as there is an option in the BIOS called “Disable Memory Error Injection”:
1576934178671.png


After setting this BIOS setting to false and enabling “ECC Injection” in MemTest86:
1576934205529.png


I ‘ve ran MemTest86 and it re-produces the issue perfectly:
1576934222731.png


As you can see, it successfully injects ECC errors, but doesn’t detect them, which is exactly the same as I was seeing when trying to trigger memory errors using unstable settings.
https://www.passmark.com/forum/memtest86/5984-how-do-you-verify-ecc-error-injection-working

Also I am very curious if this is only a Ryzen 3000 issue, as the motherboard was initially designed for Ryzen 1000 and Ryzen 2000 CPUs alone. Perhaps ECC does work for those older CPUs. Unfortunately I don’t have such a CPU to try this on (feel free to send me one for testing ).

I've forwarded this info to Asrock Rack...
 
Top